"A Republican, a Democrat, an Independent, a Libertarian, and a Green...” It sounds like the set-up to a joke. But Duluthians could have the option of voting for more than one, if a November 3 ballot measure is approved, changing the City Charter to permit “ranked-choice voting” in local (non-partisan) elections.
Ranked-choice voting allows you to vote for multiple candidates in order of preference, usually up to three, although some systems allow you to rank more or even to rank them all. The first-choice votes are tallied and, if no winner emerges with 50 percent, the candidate with the fewest votes is eliminated. If—and only if—your first-choice candidate is eliminated, your next choice is then added to the tally. This continues until either someone gets 50 percent, or until there are only two candidates left.
Ranked-choice voting (RCV) is currently used in more than a dozen US cities, including St. Paul and Minneapolis. San Francisco was the first to adopt RCV in 2002.
Proponents claim that:
•Ranked-choice voting eliminates the cost of primary elections (or runoff elections, as they’re called in California, where RCV is often called “instant runoff voting”).
•It eliminates the “spoiler effect,” in which a third-party candidate splits the vote of one major party, potentially handing a win to the other major party. If you wish to vote third-party with RCV, simply select the major party candidate as your second or third choice to avoid a spoiler.
•It reduces negative campaigning because RCV makes it more politically effective for candidates to form coalitions in the hope that voters will rank them, if not first, at least second or third. A winner-take-all mentality or attacking one’s opponents become losing strategies.
•It increases voter turnout and increases the number of citizens who run for office, both of which are beneficial for those traditionally underrepresented, such as women and racial minorities.
Critics of RCV claim that:
•Without a primary, there is not enough time to make an informed choice about every candidate on the ballot. To whatever extent RCV encourages more candidates to run, that only makes the problem worse.
•The spoiler effect still exists because voters continue to select only one candidate, despite the option of selecting more (called “undervoting”).
•RCV campaigns become such love-fests that voters can’t find out anything of substance about the candidates or even tell the difference between them.
•It reduces voter turnout and increases ballot errors in the very same underrepresented communities that were supposed to be politically served by RCV.
For most of these pros and cons, RCV simply hasn’t been around long enough to know yet. For example, cities that have adopted RCV have funneled any short-term savings into voter education, new ballot-counting software, and auditing the new system to see how it’s working.
But those are upfront expenses and a savings may eventually materialize from not holding primary elections. (Any long-term savings that might have been realized in San Francisco was counteracted by a ballot initiative in 2000 to publicly fund municipal elections.)
More importantly, every type of voting system—and there are many—all have benefits and drawbacks, according to Francis Neely, a political science professor at San Francisco State University.
Neely and his colleague, Jason McDaniel, co-authored one of the seminal studies of ranked-choice voting. “Overvoting and the Equality of Voice under Instant-Runoff Voting in San Francisco” was published in the California Journal of Politics and Policy in 2014.
“It’s hard to judge one election system from another,” says Neely. “They’re all flawed. It’s a matter of choosing which flawed system the community wants.”
Indeed, the system we have now—called “plurality voting”—has some serious problems. It’s easy to understand and use, for those of us accustomed to voting for only one candidate and the one with the most votes wins.
But in addition to the spoiler effect, this familiar method is also prone to a candidate winning by “plurality,” meaning nobody received at least 50 percent of the total votes cast. This can result in the least popular candidate prevailing, because more than 50 percent of the voters actually cast their only ballot for someone else.
The difference with RCV—for better or for worse—is illustrated by the 2010 election of former Oakland Mayor Jean Quan. Quan toppled long-time incumbent Mayor Don Perata during Oakland’s first election using RCV.
When only the first-choice ballots were counted in the first round, Perata had nearly a 10-point lead over Quan—but Perata had only 33 percent of the total ballots cast.
In the second round, after the lowest vote-getter was eliminated and those second-choice votes reallocated, Perata’s lead was a few points narrower, but he still had only 37 percent of the total vote.
In the final round, the candidate in third place was eliminated, and more of her voters had ranked Quan as their second choice, handing Quan a controversial victory at 45 percent, with 88 percent of the ballots still in play.
That’s one of the quirks of ranked-choice voting—the “exhausted ballot.” If all three of your picks are eliminated before anyone wins, your ballot is “exhausted,” meaning it’s no longer being counted in the vote totals because all your horses are out of the race.
In 2011, the potential for ballot elimination was challenged in federal court. In Dudum v. Arntz, the Ninth Circuit ruled that because “all voters are afforded a single and equal opportunity to express their preferences for three candidates,” and because no new votes are cast, an exhausted ballot does not infringe on the right to vote.
However, exhausted ballots can mean the final winner still doesn’t have 50 percent. In Oakland’s first round of counting, when all the ballots were still in play, Quan had only 24 percent of first-choice votes.
There are two ways to spin the narrative about Jean Quan’s victory. More people were at least ok with Quan for mayor, as indicated by how many ranked her second or third. She had the most votes overall and, thus, a broader mandate to lead. That’s the pro-RCV way to look at it.
On the other hand, Perata had more first-choice votes than anyone else, suggesting more enthusiasm for him than for a Johnny-come-lately whom few wanted badly enough to rank her first. She wasn’t re-elected in 2014.
“If you’re changing the voting system because you want to change the outcome? That’s a bad reason to change the system,” says Jason McDaniel, a political science professor at San Francisco State University, who co-authored the 2014 study with Francis Neely.
While McDaniel and Neely agree about their data, Neely tends to be more or less in favor of RCV. He would prefer an option that allows for ranking all the candidates, “but it’s as good a system as you’re going to get.”
McDaniel believes the benefits of high voter turnout overshadow the modern impulse to tinker with improving the way we vote. “There’s a long history of progressive reforms to get more deliberative voting. All these reforms pretty much drove voting into the gutter...The best thing we can do is make it as simple as possible to vote and let people decide for themselves.”
The Neely/McDaniel study found a higher rate of ballot errors with RCV, and the effect was most pronounced among African-Americans, Latinos, the elderly, those born outside the US, and those for whom English is a second language. One example of an RCV ballot error is giving more than one candidate the same rank, called “overvoting,” which disqualifies the ballot.
Neely points out that these same populations have higher rates of ballot error in the current system, too. In fact, the very highest error rates occur when voters are asked to select more than one candidate in a single race to fill multiple seats. “Any time you put a more complicated ballot in front of people, the error rate goes up.”
Neely also prefers the more amicable campaigning that RCV encourages. “It makes sense [for candidates] to form coalitions...‘I’m the best and everybody else stinks’ is no longer a good strategy. Just ask Don Perata.”
But McDaniel sees these extra-sweetened elections as a bug, not a feature. “If you’re trying to make the system nicer, there are going to be unintended consequences...Nobody likes negative campaigning against their own candidate, but it does increase voter information,” while RCV encourages candidates to exaggerate their similarities in the hopes of scoring a second or third ranking.
If McDaniel sounds curmudgeonly, it’s nothing compared to FairVote, a pro-RCV non-profit that has helped cities to develop ballot initiatives, to hone their new election procedures, and to educate voters—and which has a pattern of dead-agenting its critics, Scientology-style.
“[They] have relentlessly berated people (including me) who have raised concerns,” wrote Scott James for The Bay Citizen. He claims FairVote sent him “bellicose emails” after he reported on a University of San Francisco analysis of the Quan race that showed a high enough rate of disqualified ballots to potentially affect the outcome.
When McDaniel posted an online draft of a project examining the effects of RCV on voter turnout by race, FairVote went on the offense with, “A Response to [McDaniel’s] Deceptive Claims,” asserting that, “He applied sophisticated statistical tools to his limited data set and found a string of correlations...” This is less of a criticism and more like an apt description of what researchers do: Apply controls and confidence intervals to gauge the effect of variables and to determine how much their sample represents the population as a whole.
FairVote then accused McDaniel of “withdrawing” his paper when, in fact, it had been peer-reviewed and accepted for publication in The Journal of Urban Affairs.
Some closer to home have also gotten on FairVote’s bad side. In 2014, University of Minnesota political science professors Lawrence Jacobs and Joanne Miller conducted an analysis of the 2013 Minneapolis elections. Their findings are similar to Neely and McDaniel’s: Poor and minority voters had lower turnout and higher ballot errors.
Miller and Jacobs went on to publish a remarkably restrained op-ed in the Star Tribune—going out of their way to find things to praise about RCV—but: “Let’s turn down the passion...and put on our thinking caps. RCV contributes to improving our democracy in certain respects, but still falls short of its promise to improve participation by all parts of our community.”
FairVote came back swinging. “Jacobs and Miller dwell at length on the nominal difference of 1.56 percent”—a statistically significant number in the Jacobs/Miller report, once all those “sophisticated tools” had been applied.
Minneapolis City Clerk Casey Carl was appointed in 2010, after Minneapolis’ first RCV election required hand-counting the ballots. It took 15 days before all the winners were declared. The City has since purchased software that counts only the votes that come into play and maintains no record of unused votes or exhausted ballots. By 2013, most of the winners were declared on Election Night and the rest four days later.
“I’m not here to tell people which system we should use,” says Carl. “I’m supposed to be neutrally implementing the will of the people. And ranked-choice voting was the will of the people.”